scipy.stats.kstat#

scipy.stats.kstat(data, n=2, *, axis=None, nan_policy='propagate', keepdims=False)[source]#

Return the nth k-statistic (1<=n<=4 so far).

The nth k-statistic k_n is the unique symmetric unbiased estimator of the nth cumulant kappa_n.

Parameters:
dataarray_like

Input array. Note that n-D input gets flattened.

nint, {1, 2, 3, 4}, optional

Default is equal to 2.

axisint or None, default: None

If an int, the axis of the input along which to compute the statistic. The statistic of each axis-slice (e.g. row) of the input will appear in a corresponding element of the output. If None, the input will be raveled before computing the statistic.

nan_policy{‘propagate’, ‘omit’, ‘raise’}

Defines how to handle input NaNs.

  • propagate: if a NaN is present in the axis slice (e.g. row) along which the statistic is computed, the corresponding entry of the output will be NaN.

  • omit: NaNs will be omitted when performing the calculation. If insufficient data remains in the axis slice along which the statistic is computed, the corresponding entry of the output will be NaN.

  • raise: if a NaN is present, a ValueError will be raised.

keepdimsbool, default: False

If this is set to True, the axes which are reduced are left in the result as dimensions with size one. With this option, the result will broadcast correctly against the input array.

Returns:
kstatfloat

The nth k-statistic.

See also

kstatvar

Returns an unbiased estimator of the variance of the k-statistic.

moment

Returns the n-th central moment about the mean for a sample.

Notes

For a sample size n, the first few k-statistics are given by:

\[k_{1} = \mu k_{2} = \frac{n}{n-1} m_{2} k_{3} = \frac{ n^{2} } {(n-1) (n-2)} m_{3} k_{4} = \frac{ n^{2} [(n + 1)m_{4} - 3(n - 1) m^2_{2}]} {(n-1) (n-2) (n-3)}\]

where \(\mu\) is the sample mean, \(m_2\) is the sample variance, and \(m_i\) is the i-th sample central moment.

Beginning in SciPy 1.9, np.matrix inputs (not recommended for new code) are converted to np.ndarray before the calculation is performed. In this case, the output will be a scalar or np.ndarray of appropriate shape rather than a 2D np.matrix. Similarly, while masked elements of masked arrays are ignored, the output will be a scalar or np.ndarray rather than a masked array with mask=False.

References

http://mathworld.wolfram.com/k-Statistic.html

http://mathworld.wolfram.com/Cumulant.html

Examples

>>> from scipy import stats
>>> from numpy.random import default_rng
>>> rng = default_rng()

As sample size increases, n-th moment and n-th k-statistic converge to the same number (although they aren’t identical). In the case of the normal distribution, they converge to zero.

>>> for n in [2, 3, 4, 5, 6, 7]:
...     x = rng.normal(size=10**n)
...     m, k = stats.moment(x, 3), stats.kstat(x, 3)
...     print("%.3g %.3g %.3g" % (m, k, m-k))
-0.631 -0.651 0.0194  # random
0.0282 0.0283 -8.49e-05
-0.0454 -0.0454 1.36e-05
7.53e-05 7.53e-05 -2.26e-09
0.00166 0.00166 -4.99e-09
-2.88e-06 -2.88e-06 8.63e-13